findepi commented on issue #12622: URL: https://github.com/apache/datafusion/issues/12622#issuecomment-2662322613
> Furthermore, if we separate logical and physical types in LogicalPlan and PhysicalPlan, we would need to implement coercion and casting logic at both layers Casting yes (eg for constant folding) Coercion no -- "coercion" as understood at least by me, is a SQL language feature, applied by analyzer (or query builder https://github.com/apache/datafusion/issues/14618), and should not exist once we have a fully resolved plan such as physical plan (https://github.com/apache/datafusion/issues/12723) > As I previously thought about [here](https://github.com/apache/datafusion/issues/14296#issuecomment-2618344472), it may be a better approach to retain the "physical-type"—or more precisely, the "decoding-type" (arrow::DataType)—within the logical layer. This may make sense in current architecture and with current design. @jayzhan211 wdyt about https://github.com/apache/datafusion/issues/12720? This would make DF processing more flexible and more akin other systems (like e.g. Trino, Velox or Snowflake FWICT). -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org For additional commands, e-mail: github-h...@datafusion.apache.org
